INDEX

Explanations

requesting things for me

New Auto-Interp

Top Features by Cosine Similarity

Configuration

Prompts (Dashboard)

10,000 prompts, 128 tokens each

Dataset (Dashboard)

lmsys/lmsys-chat-1m

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

.nano

-0.09

æĤ¨çļĦ

-0.09

ogo

-0.09

uyo

-0.08

aim

-0.08

aea

-0.08

uldu

-0.08

US

-0.08

utin

-0.08

 sami

-0.08

POSITIVE LOGITS

me

0.61

 Ð¼Ð½Ðµ

0.51

us

0.50

ç»ĻæĪĳ

0.38

æĪĳ

0.38

 Ð¼ÐµÐ½Ñĸ

0.36

 bize

0.29

è®©æĪĳ

0.29

 tÃ´i

0.27

 Ð¼ÐµÐ½Ñı

0.26

Activations Density 0.299%